38 A.W. Phillips and R.J. Allemang 0 500 1000 1500 2000 2500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Consistency Diagram Frequency (Hz) Model Iteration cluster pole & vector pole frequency conjugate non realistic 1/condition Fig. 4.4 C-Plate consistency diagram, complex weighting, full frequency range 5 10 15 20 25 5 10 15 20 25 Mode (real) Mode (imag) MAC 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 5 10 15 20 25 5 10 15 20 25 Mode (real) Mode (imag) MAC (weighted) 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 Fig. 4.5 C-Plate, complex weighting, riMAC and riwMAC 4.5.2 C-Plate Example: Estimates with Real Weighting For this example, the entire frequency range from 200 to 2,500 Hz was again fit by the Rational Fraction Polynomial Algorithm with Z-frequency weighting (RFP-Z) in a single parameter estimation run. The final results were again determined from the CSSAMI autonomous procedure. All conditions match the modal parameter estimation process used in the last section. This time, however, instead of using the complex valued modal participation vectors as weighting vectors, real normalization of the modal participation vector (first rotated to its dominant central axis) was used to generated real-valued weighting vectors. Dramatically improved results can be observed in the following figures. It can be observed in the MAC plot (Fig. 4.7) that the coupling contamination between the 2,300 Hz repeated root modes has been eliminated. Further, it can be observed that the cross MAC between each vector and its complex conjugate is also improved. This improvement is further revealed by the
RkJQdWJsaXNoZXIy MTMzNzEzMQ==